Abstract
Background
Escherichia coli, Klebsiella spp. and Pseudomonas aeruginosa are included in the mandatory surveillance of bloodstream infections (BSI) in England.
Aim
We aimed to investigate the epidemiology of these BSIs in England April 2017–March 2022.
Methods
We extracted data on E. coli, Klebsiella spp. and P. aeruginosa BSI, categorised BSIs as healthcare-associated (HA) or community-associated (CA) and linked to antimicrobial susceptibility and mortality data. We used multivariable logistic regression models to assess predictors of mortality.
Results
The incidence rate of E. coli BSI was 71.8 per 100,000 population (95% confidence interval (CI): 71.5–72.1), Klebsiella spp. 19.1 (95% CI: 18.9–19.3) and P. aeruginosa 7.6 (95% CI: 7.5–7.7). Most (65,467/104,957) BSI episodes were CA. Case-fatality rate was 14.8% (95% CI: 14.6–14.9) in E. coli, 20.0% (95% CI: 19.6–20.3) in Klebsiella spp. and 25.8% (95% CI: 25.2–26.4) in P. aeruginosa BSI. Urinary tract infection (UTI) was the most reported primary infection for E. coli (56,961/100,834), Klebsiella spp. (9,098/22,827) and P. aeruginosa (3,204/8,484) BSI. Insertion or manipulation of urinary catheters was reported for 26.4% (16,136/61,043) of E. coli, 41.6% (4,470 /10,734) of Klebsiella spp. and 49.0% (2,127/4,341) of P. aeruginosa BSI. The adjusted odds ratio (OR) of death among hospital-onset HA-BSI compared to CA-BSI was 2.0 (95% CI: 1.9–2.2) for E. coli, 2.1 (95% CI: 2.0–2.3) for Klebsiella spp. and 1.7 (95% CI: 1.5–2.0) for P. aeruginosa.
Conclusions
Appropriate management of UTIs and urinary catheterisation is essential for reduction of these BSIs.
Keywords: Gram-negative, Blood, Antibiotics, Resistance, Case-fatality, Healthcare-associated, Regression, England
Key public health message.
What did you want to address in this study and why?
Bloodstream infections (BSI) caused by bacteria of Escherichia coli, Klebsiella species and Pseudomonas aeruginosa can be hard to treat and prolong a patient’s stay in hospital. We describe the spread of these infections in England, the most common factors among cases, and percentage of cases where a patient died between April 2017 and March 2022.
What have we learnt from this study?
Bloodstream infections were more common in older patients. A greater percentage of Klebsiella species and P. aeruginosa BSI were healthcare-associated cases and ended in death compared with E. coli BSI. Urinary tract infections were the most common primary focus of infection, and among these, patients from about half of Klebsiella species and P. aeruginosa BSI cases had a urinary catheter inserted or manipulated.
What are the implications of your findings for public health?
Despite some similarities between these infections, different strategies are likely to be required to reduce infection rates. Focusing on older patients with UTI from community settings is likely to yield the greatest reductions in E. coli BSI. However, for Klebsiella species and P. aeruginosa BSI, greater reductions may be achieved by focusing on patients that have had a urinary catheter inserted or manipulated.
Introduction
Between April 2022 and March 2023, Escherichia coli, Klebsiella spp. and Pseudomonas aeruginosa accounted for 75% of Gram-negative bloodstream infections (GNBSI) and 38% of all bloodstream infections (BSI), excluding those caused by coagulase-negative Staphylococci, in England, Wales and Northern Ireland [1]. Among these, Klebsiella spp. and P. aeruginosa BSI are more likely to be healthcare-associated (HA), impact patients with weakened immune systems [2-4] and result in high mortality. Consequently, GNBSI lead to lengthy hospital stays and increased costs of care [5-7], thereby constituting a disease burden on patients and health services. Naylor et al. 2019 estimated the annual costs of E. coli bacteraemia in England to be Great British Pound (GBP) 14,346,400 (year 2012 GBP value) with an excess cost of GBP 420 (95% confidence interval (CI): GBP 220–630) per patient when associated with third-generation cephalosporin resistance. The burden of these infections on the United Kingdom (UK) population and healthcare services contributed to the ambition of the UK government to halve the number of HA-GNBSI patients identified between April 2016 and March 2017 by 2023–24 [8]. To support this initiative, in April 2017, the mandatory surveillance of E. coli BSI in England was expanded to include Klebsiella spp. and P. aeruginosa BSI.
National Health Service (NHS) acute trusts (secondary care providers) notify all laboratory-confirmed cases of E. coli, Klebsiella spp. and P. aeruginosa BSI to the Healthcare-associated infections (HCAI) Data Capture System (HCAI DCS) of the United Kingdom Health Security Agency (UKHSA). Post-mortem samples are not included. Blood samples are usually collected from patients with symptoms such as persisting fever and abnormal heart rate in line with clinical standards set by NHS England and tested for antimicrobial susceptibility using antimicrobial testing panels based on local recommendations and prescribing policy. In a survey in 2018, 95% of responding laboratories in England (survey response rate 94%) indicated that the European Committee on Antimicrobial Susceptibility Testing (EUCAST; https://www.eucast.org/) or the British Society for Antimicrobial Chemotherapy (BSAC; https://bsac.org.uk/) susceptibility testing breakpoints were used with some variations in testing methods.
In England, between April 2012 and March 2014, genital or urinary tract was the most common primary focus of E. coli BSI (41.1%) and 76.1% of the infections were community-onset [9]. In the sentinel study of Abernethy et al. 2017, half of the patients had previous healthcare exposure in the month before the bacteraemia, with antimicrobial therapy and urinary catheterisation reported as common treatments [10]. There are no studies on the mandatory surveillance of Klebsiella spp. and P. aeruginosa BSI in England.
We aimed to present findings from the first 5 years of the expanded GNBSI national surveillance in England, providing key information on the epidemiology of these infections, and to support the development of infection prevention and control strategies in the UK. In this study, we report the incidence, predisposing risk factors, antimicrobial resistance and 30-day all-cause case-fatality rates and mortality rates associated with these infections.
Methods
Data collection
Gram-negative bloodstream infections
On 17 October 2022, we extracted data from the UKHSA HCAI DCS [11] on E. coli, Klebsiella spp. and P. aeruginosa BSIs diagnosed by NHS acute trusts in England between 1 April 2017 and 31 March 2022. Data on patient demographics, case and admission details, treatment specialty, predisposing risk factors within 28 days before the specimen date and the suspected primary focus of the infection were extracted. Information on the primary focus of infection is voluntary to include in the report. Multiple risk factors can be reported with each diagnosis. The reported risk factors depend on the primary focus of infection. For example, if the primary focus of infection is a respiratory tract infection, then whether (or not) the patient was intubated or extubated earlier will be reported but not whether the patient had a urinary catheter inserted or manipulated. The opposite is the case when the primary focus of infection is the urinary tract. The applicable risk factors for each primary focus of infection are presented in Supplementary Table 1.
A BSI episode was defined as the first report of a GNBSI from a patient within a fixed 14-day period. Subsequent reports within this fixed period were considered duplicate reports and excluded from all analyses. Episodes of E. coli, Klebsiella spp. and P. aeruginosa BSI were defined separately meaning that a patient could have an episode of any of these pathogens on the same day. Furthermore, the fixed 14-day episode length meant that a patient could have another BSI episode between the 15th and 29th day after the previous BSI episodes. More details can be seen in Supplementary Table 2.
Bloodstream infections were categorised as HA or community-associated (CA). A HA-BSI could be either hospital-onset (HOHA) or community-onset HA (COHA). A HOHA-BSI was one where a blood sample with detection of E. coli, Klebsiella spp. or P. aeruginosa was taken ≥ 3 days after an admission to an NHS acute trust (Box). A COHA-BSI was one where the patient was previously discharged from the same NHS acute trust ≤ 28 days before the specimen date. Any BSI not defined as a HA-BSI was considered a CA-BSI. Information to determine a HA-BSI was only available for BSI episodes after April 2020. Therefore, analyses involving healthcare-association were limited to the period between April 2020 and March 2022.
Box. Definition of healthcare-associated and community-associated bloodstream infections caused by Escherichia coli, Klebsiella species and Pseudomonas aeruginosa, England, April 2020–March 2022.
Bloodstream infection (BSI):
• Presence of viable microorganisms in the bloodstream, in this study detection of Escherichia coli, Klebsiella spp. or Pseudomonas aeruginosa from a blood specimen.
Healthcare-associated BSI (HA-BSI):
• BSI associated with a healthcare visit.
• Hospital-onset (HOHA):
o Blood specimen taken ≥ 3 days after an admission to an NHS acute trust.
• Community-onset (COHA):
o BSI is not a HOHA-BSI and patient was previously discharged from the same NHS acute trust ≤ 28 days before the specimen date.
Community-associated BSI (CA-BSI):
• Any BSI that was not HA BSI.
Mortality
Information on deaths and dates of deaths of the patients with BSIs were derived by deterministic linkage using the patient NHS numbers (unique patient identifier) and dates of birth, to a central repository of patient information (NHS SPINE). We analysed mortality as 30-day all-cause deaths, i.e. deaths from any cause within 30 days of the specimen date. Since a BSI’s episode length was 14 days, a patient could have up to three BSI episodes within this 30-day period. To prevent overestimating the number of deaths, only the most recent BSI episode of each pathogen before the date of death (or the most recent episode of each pathogen for unlinked patients) within the 30-day period was included in all mortality analysis. More details can be seen in Supplementary Table 2.
Antimicrobial susceptibility results
Susceptibility results for gentamicin, amikacin, netilmicin, tobramycin, imipenem, cefotaxime ceftazidime, piperacillin-tazobactam, ciprofloxacin and amoxicillin-clavulanic acid were derived by deterministic linkage using the patient’s NHS number and date of birth between the BSI episodes and data from UKHSA’s voluntary laboratory surveillance database Second Generation Surveillance System. If there were different susceptibility results for each antimicrobial within 14 days of the BSI episodes, the most resistant result per antimicrobial was assigned to the episode.
Population estimates
England’s mid-year population estimates by age, sex and region for 2017 to 2022 were downloaded from the Office for National Statistics (ONS) website [12], while the quarterly average number of overnight occupied beds and day admissions were sourced from NHS England’s published Bed Availability and Occupancy (KH03) returns [13].
Data analysis
Incidence and mortality rates
Incidence and mortality rates were calculated using the total number of BSI episodes and 30-day all-cause deaths respectively, against England’s mid-year population for all reported episodes or total overnight bed-days and day admissions for HA-BSI. Case-fatality rates (CFR) per pathogen were calculated as the percentage of 30-day all-cause deaths among all BSI episodes included in the mortality analyses. Regional analysis was based on seven NHS administrative regions covering England.
Antimicrobial resistance
Antimicrobial resistance was calculated as the percentage of all BSI episodes with susceptibility results that have been assigned resistant (R). This was calculated separately for each antimicrobial or antimicrobial class. In line with EUCAST definitions [14], we considered susceptible (S) and intermediate (I) results (now defined as susceptible with increased exposure) as susceptible. Aminoglycoside resistance was defined as resistance to either gentamicin, amikacin, netilmicin or tobramycin. For E. coli and Klebsiella spp., carbapenem resistance was defined as resistance to either imipenem or meropenem. For P. aeruginosa isolates, carbapenem resistance was defined as resistance to meropenem only. However, we used the result of testing against imipenem as a proxy if there was no susceptibility result for meropenem. Third-generation cephalosporin resistance was defined as resistance to cefotaxime or ceftazidime for E. coli and Klebsiella spp. isolates but for P. aeruginosa isolates, as resistance to ceftazidime only. Multidrug resistance (MDR) for E. coli and Klebsiella spp. isolates was defined as resistance to any three of the antimicrobial agents or classes: aminoglycosides, piperacillin-tazobactam, third-generation cephalosporin, carbapenems, ciprofloxacin and amoxicillin-clavulanic acid. For isolates of P. aeruginosa, susceptibility to amoxicillin-clavulanic acid was not included due to expected intrinsic resistance.
Predictors of mortality
Multivariable logistic regression models were used to assess HA-BSI (HOHA and COHA separately) and MDR for increased risk of mortality, expressed as adjusted odds ratios (aOR). Each model was adjusted for patients’ age group, sex, geographical region, year and month of infection as confounding variables. Episodes where data on the outcome and explanatory variables could not be determined were excluded from the regression analysis. We tested for multicollinearity among these variables by calculating scaled generalised variance inflation factors.
Statistical significance was defined as p values ≤ 0.05. Fisher’s exact test p values and Wald 95% confidence intervals (CI) were reported for relative risks (RR). Wald p values and 95% CIs were reported for rate ratios. Wald p values and profile likelihood ratio 95% CIs were reported for the aORs. Scaled GVIF scores of between 1 and 1.5 were interpreted as no considered evidence of multicollinearity. All analyses were done in R version 4.0.2 (https://www.r-project.org/) with the tidyverse, epitools, stats, car and diyar packages [15,16].
Results
Gram-negative bloodstream infections
After removing duplicate reports as described in the method section, we identified 277,177 GNBSI episodes among 240,796 patients with BSI between 1 April 2017 and 31 March 2022. There were 202,054 E. coli, 53,751 Klebsiella spp. and 21,372 P. aeruginosa episodes (Table 1). The mean age of the patients was similar across all three pathogens: from 70.4 years (standard deviation (SD): 19.3) in E. coli to 67.2 years (SD: 20.8) in P. aeruginosa. Male patients had 99,512 (49.3%) episodes of E. coli, 33,638 (62.6%) Klebsiella spp. and 13,735 (64.3%) P. aeruginosa. The overall incidence rates of BSI were 71.8 episodes per 100,000 population (95% CI: 71.5–72.1) for E. coli, 19.1 (95% CI: 18.9–19.3) for Klebsiella spp. and 7.6 (95% CI: 7.5–7.7) for P. aeruginosa (Table 1). The annual incidence rate of E. coli BSI declined from 73.6 (95% CI: 72.9–74.4) in 2017–18 to 67.1 (95% CI: 66.4–67.7) in 2021–22. Conversely, over the same period, the incidence rate of Klebsiella spp. BSI increased from 17.5 (95% CI: 17.1–17.8) to 20.1 (95% CI: 19.7–20.4). The incidence rate of P. aeruginosa BSI was 7.7 (95% CI: 7.5–7.9) in 2017–18 and 7.6 (95% CI: 7.4–7.9) in 2021–22.
Table 1. Frequency and incidence rates of Escherichia coli (n = 202,054), Klebsiella species (n = 53,751) and Pseudomonas aeruginosa (n = 21,372) bloodstream infections, England, 1 April 2017–31 March 2022 (n = 240,796 patients)a .
| Characteristics | Escherichia coli | Klebsiella species | Pseudomonas aeruginosa | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | % | IRb | 95% CI | n | % | IR | 95% CI | n | % | IR | 95% CI | |
| Number of episodes | ||||||||||||
| April 2017–March 2018 | 41,020 | 20.3 | 73.6 | 72.9–74.4 | 9,725 | 18.1 | 17.5 | 17.1–17.8 | 4,290 | 20.1 | 7.7 | 7.5–7.9 |
| April 2018–March 2019 | 43,150 | 21.4 | 77.0 | 76.3–77.7 | 10,632 | 19.8 | 19.0 | 18.6–19.3 | 4,173 | 19.5 | 7.4 | 7.2–7.7 |
| April 2019–March 2020 | 43,278 | 21.4 | 76.6 | 75.9–77.4 | 10,988 | 20.4 | 19.5 | 19.1–19.8 | 4,314 | 20.2 | 7.6 | 7.4–7.9 |
| April 2020–March 2021 | 36,683 | 18.2 | 64.9 | 64.2–65.5 | 11,065 | 20.6 | 19.6 | 19.2–19.9 | 4,275 | 20.0 | 7.6 | 7.3–7.8 |
| April 2021–March 2022 | 37,923 | 18.8 | 67.1 | 66.4–67.7 | 11,341 | 21.1 | 20.1 | 19.7–20.4 | 4,320 | 20.2 | 7.6 | 7.4–7.9 |
| Total | 202,054 | 71.8 | 71.5–72.1 | 53,751 | 19.1 | 18.9–19.3 | 21,372 | 7.6 | 7.5–7.7 | |||
| Age of females (years) | ||||||||||||
| < 1 | 1,365 | 1.3 | 90.4 | 85.6–95.3 | 449 | 2.2 | 29.7 | 27.0–32.6 | 128 | 1.7 | 8.5 | 7.1–10.1 |
| 1–14 | 578 | < 1 | 2.5 | 2.3–2.7 | 300 | 1.5 | 1.3 | 1.1–1.4 | 258 | 3.4 | 1.1 | 1.0–1.3 |
| 15–44 | 10,997 | 10.7 | 20.8 | 20.4–21.2 | 2,138 | 10.6 | 4.0 | 3.9–4.2 | 809 | 10.6 | 1.5 | 1.4–1.6 |
| 45–64 | 18,454 | 18.0 | 50.6 | 49.8–51.3 | 4,918 | 24.5 | 13.5 | 13.1–13.9 | 1,806 | 23.7 | 4.9 | 4.7–5.2 |
| 65–74 | 18,957 | 18.5 | 131.2 | 129.3–133.0 | 4,348 | 21.6 | 30.1 | 29.2–31.0 | 1,725 | 22.6 | 11.9 | 11.4–12.5 |
| 75–84 | 27,159 | 26.5 | 295.6 | 292.1–299.1 | 4,624 | 23.0 | 50.3 | 48.9–51.8 | 1,614 | 21.2 | 17.6 | 16.7–18.4 |
| ≥ 85 | 24,913 | 24.3 | 569.1 | 562.1–576.2 | 3,310 | 16.5 | 75.6 | 73.1–78.2 | 1,291 | 16.9 | 29.5 | 27.9–31.1 |
| Total | 102,423 | 50.7 | 72.0 | 71.6–72.5 | 20,087 | 37.4 | 14.1 | 13.9–14.3 | 7,631 | 35.7 | 5.4 | 5.2–5.5 |
| Age of males (years) | ||||||||||||
| < 1 | 2,104 | 2.1 | 132.1 | 126.6–137.9 | 600 | 1.8 | 37.7 | 34.7–40.8 | 159 | 1.2 | 10.0 | 8.5–11.7 |
| 1–14 | 478 | < 1 | 2.0 | 1.8–2.1 | 401 | 1.2 | 1.6 | 1.5–1.8 | 313 | 2.3 | 1.3 | 1.1–1.4 |
| 15–44 | 3,857 | 3.9 | 7.2 | 6.9–7.4 | 1,976 | 5.9 | 3.7 | 3.5–3.8 | 942 | 6.9 | 1.7 | 1.6–1.9 |
| 45–64 | 18,341 | 18.4 | 51.7 | 50.9–52.4 | 7,453 | 22.2 | 21.0 | 20.5–21.5 | 2,834 | 20.6 | 8.0 | 7.7–8.3 |
| 65–74 | 23,038 | 23.2 | 171.9 | 169.7–174.2 | 7,876 | 23.4 | 58.8 | 57.5–60.1 | 3,171 | 23.1 | 23.7 | 22.8–24.5 |
| 75–84 | 30,119 | 30.3 | 394.4 | 390.0–398.9 | 9,309 | 27.7 | 121.9 | 119.4–124.4 | 3,746 | 27.3 | 49.1 | 47.5–50.7 |
| ≥ 85 | 21,575 | 21.7 | 840.9 | 829.7–852.2 | 6,023 | 17.9 | 234.7 | 228.8–240.7 | 2,570 | 18.7 | 100.2 | 96.3–104.1 |
| Total | 99,512 | 49.3 | 71.5 | 71.1–72.0 | 33,638 | 62.6 | 24.2 | 23.9–24.4 | 13,735 | 64.3 | 9.9 | 9.7–10.0 |
| Healthcare-associationc | n = 74,094 | n = 22,312 | n = 8,551 | |||||||||
| Healthcare-associated | 24,066 | 32.5 | 13.2 | 13.0–13.4 | 10,568 | 47.4 | 5.8 | 5.7–5.9 | 4,856 | 56.8 | 2.7 | 2.6–2.7 |
| Community-associated | 50,028 | 67.5 | 17.8 | 17.6–17.9 | 11,744 | 52.6 | 4.2 | 4.1–4.3 | 3,695 | 43.2 | 1.3 | 1.3–1.4 |
| Primary focus of infection | n = 100,834 | n = 22,827 | n = 8,484 | |||||||||
| Urinary tract | 56,961 | 56.5 | NA | 9,098 | 39.9 | NA | 3,204 | 37.8 | NA | |||
| Hepatobiliary | 19,847 | 19.7 | 5,244 | 23.0 | 533 | 6.3 | ||||||
| Gastrointestinal (not hepatobiliary) | 8,109 | 8.0 | 2,117 | 9.3 | 687 | 8.1 | ||||||
| Respiratory tract | 7,310 | 7.2 | 2,879 | 12.6 | 1,523 | 18.0 | ||||||
| Other | 5,699 | 5.7 | 2,698 | 11.8 | 2,136 | 25.2 | ||||||
| No underlying focus of infection | 2,908 | 2.9 | 791 | 3.5 | 401 | 4.7 | ||||||
BSI: bloodstream infection; CI: confidence interval; IR: incidence rate: NA: not applicable.
a A patient could have more than one BSI episode of infection.
b Incidence rate of healthcare-associated BSI is per 100,000 overnight bed-days and day admissions. Incidence rate of hospital-onset BSI is per overnight bed-days. Every other incidence rate is per 100,000 population.
c Only data on infections between April 2020 and March 2022 were included.
Demographic distribution
In patients aged ≥ 45 years, incidence rates were higher in males with Klebsiella spp. and P. aeruginosa BSI (Table 1) than in females. In male patients aged ≥ 85 years, the incidence rate ratio (IRR) compared to females with E. coli was 1.48 (95% CI: 1.45–1.50, p < 0.01), while IRR for Klebsiella spp. was 3.1 (95% CI: 3.0–3.2; p < 0.01) and 3.4 for P. aeruginosa (95% CI: 3.2–3.6; p < 0.01). The incidence rate of E. coli BSI in females aged 15–44 years was higher than in males (IRR = 2.9; 95% CI: 2.8–3.0; p < 0.01). The geographical distribution of incidence rates of HA-BSIs and CA-BSIs between 2020–21 and 2021–22 is presented in Supplementary Table 5.
Healthcare-association
Between 1 April 2020 and 31 March 2022, most (62.4%; 65,467/104,957) E. coli, Klebsiella spp. and P. aeruginosa BSI episodes were CA. However, 32.5% (24,066/74,094) of E. coli, 47.4% (10,568/22,312) of Klebsiella spp. and 56.8% (4,856/8,551) of P. aeruginosa BSI episodes were HA (Table 1). The incidence rate for HA-BSI of E. coli, Klebsiella spp. and P. aeruginosa was 13.2 (95% CI: 13.0–13.4), 5.8 (95% CI: 5.7–5.9) and 2.7 (95% CI: 2.6–2.7) episodes per 100,000 overnight bed-days and day admissions, respectively. Of the CA-BSIs, 6.8% (n = 7,165) had a predisposing risk factor based on the reported primary focus of infection: 6.8% of E. coli (n = 4,970), 6.8% of Klebsiella spp. (n = 1,512) and 8% of P. aeruginosa (n = 683). The predisposing risk factors are detailed in Supplementary Table 1.
Primary focus of infection and predisposing risk factors
When information was available among GNBSI episodes between 1 April 2017 and 31 March 2022, urinary tract infections were the most reported primary infections: 56.5% (56,961/100,834) of E. coli, 39.9% (9,098/22,827) of Klebsiella spp. and 37.8% (3,204/8,484) of P. aeruginosa BSI episodes. Respiratory tract infection (RTI) was more common in P. aeruginosa (18.0%; 1,523/8,484) than in E. coli (7.2%; 7,310/100,834) or Klebsiella spp. (12.6%; 2,879/22,827) BSI. More information can be seen in Supplementary Figures 1-2 and Supplementary Table 4.
Among all 277,177 GNBSI episodes, 141,131 (50.9%) were reported with information on predisposing risk factors within 28 days prior to the episode. The most reported risk factors based on the primary focus of infection were the insertion or manipulation of urinary catheters or vascular devices. Information on insertion or manipulation of urinary catheters was reported for 26.4% (16,136/61,043) of E. coli, 41.6% (4,470/10,734) of Klebsiella spp. and 49.0% (2,127/4,341) of P. aeruginosa episodes. Vascular devices were inserted in 27.4% (3,926/14,324) E. coli, 49.1% (2,224/4,526) Klebsiella spp. and 52.8% (1,266/2,396) P. aeruginosa episodes. During the study period, information on urinary catheterisation was known for 72% of E. coli, 62% of Klebsiella spp. and 55% of P. aeruginosa BSI with a UTI primary focus of infection (or none at all).
Antimicrobial resistance
Between 1 April 2020 and 31 March 2022, we received AMR data for 90.4% (67,465/74,606) of E. coli isolates, 89.3% (20,008/22,406) of Klebsiella spp. isolates and 89.7% (7,711/8,595) of P. aeruginosa isolates. Of the P. aeruginosa isolates from patients, 6.1% (461/7,558; 95% CI: 5.6–6.7%) were carbapenem resistant compared to < 1% of Klebsiella spp. and E. coli isolates (Table 2). Resistance to amoxicillin-clavulanic acid was 41.9% (27,135/64,693; 95% CI: 41.6–42.3%) in E. coli isolates and 29.9% (5,754/19,235; 95% CI: 29.3–30.6%) in Klebsiella spp. isolates. Resistance to third-generation cephalosporins was 11.6% (6,546/56,657; 95% CI: 11.3–11.8%) in E. coli, 13.4% (2,268/16,897; 95% CI: 12.9–13.9%) in Klebsiella spp. and 6.6% (497/7,494; 95% CI: 6.1–7.2%) in P. aeruginosa isolates. Resistance to piperacillin-tazobactam was 9.6% (6,007/62,749; 95% CI: 9.3–9.8%) in E. coli, 16.3% (3,063/18,779; 95% CI: 15.8–16.8%) in Klebsiella spp. and 8.0% (580/7,287; 95% CI: 7.3–8.6%) in P. aeruginosa isolates. Antimicrobial resistance was generally higher in isolates from HA-BSI cases compared with those from CA-BSI cases (Table 2). Of the 66,617 tested isolates of E. coli, 10.5% (n = 7,003; 95% CI: 10.3–10.7) were MDR, versus 11.6% (2,294/19,780; 95% CI: 11.2–12.1) of Klebsiella spp. isolates and 3.6% (276/7,608; 95% CI: 3.2–4.1) of P. aeruginosa isolates. Isolates from HA-BSI were more resistant. For example, 14.9% (3,215/21,582; 95% CI: 14.4–15.4) of E. coli, 16.4% (1,528/9,338; 15.6–17.1) of Klebsiella spp. and 5.2% (225/4,296; 95% CI: 4.6–5.9) of P. aeruginosa isolates were MDR.
Table 2. Antimicrobial resistance in Escherichia coli (n = 67,465), Klebsiella species (n = 20,008) and Pseudomonas aeruginosa (n = 7,711) bloodstream infections in England, by healthcare or community association, England, 1 April 2020–31 March 2022a .
| Healthcare-association | Escherichia coli | Klebsiella species | Pseudomonas aeruginosa | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Tested | Resistant | Tested | Resistant | Tested | Resistant | |||||||||
| n | % | % | 95% CI | n | % | % | 95% CI | n | % | % | 95% CI | |||
| Amoxicillin-clavulanic acid | ||||||||||||||
| Total | 64,693 | 86.7 | 41.9 | 41.6–42.3 | 19,235 | 85.8 | 29.9 | 29.3–30.6 | NA | NA | ||||
| HA | 20,904 | 86.9 | 49.8 | 49.1–50.5 | 9,059 | 85.7 | 37.5 | 36.5–38.5 | ||||||
| CA | 43,333 | 86.6 | 38.2 | 37.7–38.6 | 10,093 | 85.9 | 23.1 | 22.3–24.0 | ||||||
| Aminoglycosides | ||||||||||||||
| Total | 67,113 | 90.0 | 11.2 | 11.0–11.5 | 19,895 | 88.8 | 7.9 | 7.5–8.3 | 7,291 | 84.8 | 4.2 | 3.7–4.7 | ||
| HA | 21,701 | 90.2 | 15.1 | 14.7–15.6 | 9,362 | 88.6 | 10.3 | 9.7–10.9 | 4,124 | 84.9 | 4.7 | 4.0–5.3 | ||
| CA | 44,948 | 89.8 | 9.4 | 9.1–9.6 | 10,449 | 89.0 | 5.7 | 5.3–6.2 | 3,130 | 84.7 | 3.6 | 3.0–4.3 | ||
| Ciprofloxacin | ||||||||||||||
| Total | 65,552 | 87.9 | 17.3 | 17.0–17.6 | 19,442 | 86.8 | 11.3 | 10.9–11.8 | 7,593 | 88.3 | 7.4 | 6.8–8.0 | ||
| HA | 21,207 | 88.1 | 22.3 | 21.7–22.8 | 9,169 | 86.8 | 14.7 | 14.0–15.4 | 4,293 | 88.4 | 8.9 | 8.1–9.8 | ||
| CA | 43,886 | 87.7 | 14.8 | 14.5–15.2 | 10,191 | 86.8 | 8.3 | 7.8–8.9 | 3,263 | 88.3 | 5.5 | 4.7–6.3 | ||
| Carbapenemsb | ||||||||||||||
| Total | 65,997 | 88.5 | 0.2 | 0.1–0.2 | 19,413 | 86.6 | 0.6 | 0.5–0.7 | 7,558 | 87.9 | 6.1 | 5.6–6.7 | ||
| HA | 21,401 | 88.9 | 0.2 | 0.2–0.3 | 9,145 | 86.5 | 1.0 | 0.8–1.3 | 4,271 | 88.0 | 8.8 | 8.0–9.7 | ||
| CA | 44,141 | 88.2 | 0.2 | 0.1–0.2 | 10,185 | 86.7 | 0.3 | 0.2–0.4 | 3,249 | 87.9 | 2.5 | 2.0–3.1 | ||
| Third-generation cephalosporinsc | ||||||||||||||
| Total | 56,657 | 75.9 | 11.6 | 11.3–11.8 | 16,897 | 75.4 | 13.4 | 12.9–13.9 | 7,494 | 87.2 | 6.6 | 6.1–7.2 | ||
| HA | 18,187 | 75.6 | 15.4 | 14.9–16.0 | 7,952 | 75.2 | 18.6 | 17.8–19.5 | 4,245 | 87.4 | 9.3 | 8.4–10.2 | ||
| CA | 38,042 | 76.0 | 9.7 | 9.4–10.0 | 8,868 | 75.5 | 8.8 | 8.2–9.4 | 3,212 | 86.9 | 3.1 | 2.6–3.8 | ||
| Piperacillin–tazobactam | ||||||||||||||
| Total | 62,749 | 84.1 | 9.6 | 9.3–9.8 | 18,779 | 83.8 | 16.3 | 15.8–16.8 | 7,287 | 84.8 | 8.0 | 7.3–8.6 | ||
| HA | 20,434 | 84.9 | 13.3 | 12.9–13.8 | 8,893 | 84.2 | 22.2 | 21.3–23.1 | 4,138 | 85.2 | 10.9 | 10.0–11.9 | ||
| CA | 41,860 | 83.7 | 7.7 | 7.5–8.0 | 9,806 | 83.5 | 11.0 | 10.4–11.7 | 3,113 | 84.2 | 4.0 | 3.4–4.8 | ||
| Multidrug resistanced | ||||||||||||||
| Total | 66,617 | 89.3 | 10.5 | 10.3–10.7 | 19,780 | 88.3 | 11.6 | 11.2–12.1 | 7,608 | 88.5 | 3.6 | 3.2–4.1 | ||
| HA | 21,582 | 89.7 | 14.9 | 14.4–15.4 | 9,338 | 88.4 | 16.4 | 15.6–17.1 | 4,296 | 88.5 | 5.2 | 4.6–5.9 | ||
| CA | 44,575 | 89.1 | 8.4 | 8.1–8.6 | 10,359 | 88.2 | 7.3 | 6.8–7.9 | 3,275 | 88.6 | 1.5 | 1.1–2.0 | ||
BSI: bloodstream infection; CA: community-associated bloodstream infection; CI: confidence interval; E.: Escherichia; HA: healthcare-associated bloodstream infection; NA: not applicable; P.: Pseudomonas.
a Includes 466 E. coli, 84 Klebsiella spp. and 38 P. aeruginosa BSI cases that could not be classified as CA or HA due to missing data on discharge prior to the case.
b For P. aeruginosa isolates, carbapenem resistance was defined as resistance to meropenem only however, an imipenem result was used as a proxy if there was no susceptibility result for meropenem.
c Third-generation cephalosporin resistance was defined as resistance to cefotaxime or ceftazidime for E. coli and Klebsiella spp. isolates but was defined as resistance to ceftazidime only for P. aeruginosa isolates.
d Multidrug resistance (MDR) for E. coli and Klebsiella spp. isolates was defined as resistance to any three of the following antimicrobial agents or classes: aminoglycosides, piperacillin-tazobactam, third-generation cephalosporin, carbapenems, ciprofloxacin and amoxicillin-clavulanic acid. For P. aeruginosa, the same definition was applied, except for susceptibility to amoxicillin-clavulanic acid which was not considered.
Mortality and case-fatality
Information on patient death was not known for 7,561 (2.7%) of 277,177 episodes. Additionally, 1,450 episodes within the 30-day period for all-cause deaths were excluded from the analyses to avoid over reporting the number of deaths. Case-fatality rate of E. coli BSI was 14.8% (29,036/196,384; 95% CI: 14.6–14.9%), lower than Klebsiella spp. or P. aeruginosa BSI (Table 3). However, the highest mortality rate (10.3 deaths per 100,000 population-years; 95% CI: 10.2–10.4) was in E. coli BSI (Table 3). Conversely, CFR for P. aeruginosa BSI was higher (5,347/20,755; 25.8%; 95% CI: 25.2–26.4%) but the mortality rate was lower (1.9 deaths per 100,000 population-years; 95% CI: 1.9–2.0%). Among patients aged ≥ 15 years with E. coli and Klebsiella spp. BSI, CFR increased with age (Table 3). The largest differences in CFRs between sexes were in patients with E. coli BSI who were aged 15–44 years (RR = 3.5; 95% CI: 3.0–4.1; p < 0.01). The RRs of CFR in HA-BSI compared with CA-BSI were 1.6 (95% CI: 1.5–1.6; p < 0.01) for E. coli, 1.4 (95% CI: 1.4–1.5; p < 0.01) for Klebsiella spp. and 1.2 (95% CI: 1.1–1.3; p < 0.01) for P. aeruginosa. The RR of CFR in MDR-BSI compared with non-MDR-BSI was 1.2 (95% CI: 1.1–1.2; p < 0.01) for E. coli, 1.1 (95% CI: 1.0–1.2; p < 0.01) for Klebsiella spp. and 1.0 (RR: 1.0; 95% CI; 0.9–1.3: p = 0.7) for P. aeruginosa BSI cases.
Table 3. Case-fatality rate and mortality rates of Escherichia coli (n = 29,036), Klebsiella spp. (n = 10,396). and Pseudomonas aeruginosa (n = 5,347) bloodstream infections within 30 days of sampling, England, 1 April 2017–31 March 2022.
| Characteristics | Escherichia coli | Klebsiella species | Pseudomonas aeruginosa | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | MRa | 95% CI | CFR% | 95% CI | n | MR | 95% CI | CFR% | 95% CI | n | MR | 95% CI | CFR% | 95% CI | |
| Total deaths | 29,036 | 10.3 | 10.2–10.4 | 14.8 | 14.6–14.9 | 10,396 | 3.7 | 3.6–3.8 | 20.0 | 19.6–20.3 | 5,347 | 1.9 | 1.9–2.0 | 25.8 | 25.2–26.4 |
| Number of deaths | |||||||||||||||
| April 2017–March 2018 | 5,887 | 10.6 | 10.3–10.8 | 14.8 | 14.5–15.2 | 1,896 | 3.4 | 3.3–3.6 | 20.3 | 19.5–21.1 | 1,121 | 2.0 | 1.9–2.1 | 27.0 | 25.7–28.4 |
| April 2018–March 2019 | 5,825 | 10.4 | 10.1–10.7 | 13.9 | 13.6–14.2 | 1,934 | 3.5 | 3.3–3.6 | 18.8 | 18.0–19.5 | 977 | 1.7 | 1.6–1.9 | 24.2 | 22.9–25.5 |
| April 2019–March 2020 | 6,017 | 10.7 | 10.4–10.9 | 14.3 | 14.0–14.7 | 2,048 | 3.6 | 3.5–3.8 | 19.2 | 18.5–20.0 | 1,034 | 1.8 | 1.7–1.9 | 24.7 | 23.4–26.0 |
| April 2020–March 2021 | 5,758 | 10.2 | 9.9–10.4 | 16.1 | 15.7–16.5 | 2,357 | 4.2 | 4.0–4.3 | 22.0 | 21.2–22.7 | 1,159 | 2.0 | 1.9–2.2 | 27.8 | 26.4–29.2 |
| April 2021–March 2022 | 5,549 | 9.8 | 9.6–10.1 | 15.0 | 14.7–15.4 | 2,161 | 3.8 | 3.7–4.0 | 19.6 | 18.9–20.4 | 1,056 | 1.9 | 1.8–2.0 | 25.1 | 23.8–26.4 |
| Total | 29,036 | 10.3 | 10.2–10.4 | 14.8 | 14.6–14.9 | 10,396 | 3.7 | 3.6–3.8 | 20.0 | 19.6–20.3 | 5,347 | 1.9 | 1.9–2.0 | 25.8 | 25.2–26.4 |
| Age of females (years) | |||||||||||||||
| < 1 | 87 | 5.8 | 4.6–7.1 | 8.4 | 6.8–10.3 | 39 | 2.6 | 1.8–3.5 | 10.2 | 7.3–13.6 | 25 | 1.7 | 1.1–2.4 | 25.0 | 16.9–34.7 |
| 1–14 | 15 | 0.1 | < 0.1–0.1 | 2.8 | 1.6–4.6 | 4 | < 0.1 | < 0.1– < 0.1 | 1.4 | < 1–3.7 | 15 | 0.1 | < 0.1–0.1 | 6.1 | 3.5–9.9 |
| 15–44 | 280 | 0.5 | 0.5–0.6 | 2.7 | 2.4–3.0 | 167 | 0.3 | 0.3–0.4 | 8.1 | 7.0–9.4 | 95 | 0.2 | 0.1–0.2 | 12.2 | 10.0–14.7 |
| 45–64 | 1,842 | 5.0 | 4.8–5.3 | 10.3 | 9.9–10.7 | 794 | 2.2 | 2.0–2.3 | 16.6 | 15.6–17.7 | 430 | 1.2 | 1.1–1.3 | 24.7 | 22.6–26.7 |
| 65–74 | 2,268 | 15.7 | 15.1–16.4 | 12.3 | 11.8–12.8 | 889 | 6.2 | 5.8–6.6 | 20.9 | 19.7–22.2 | 532 | 3.7 | 3.4–4.0 | 31.5 | 29.3–33.8 |
| 75–84 | 3,774 | 41.1 | 39.8–42.4 | 14.2 | 13.8–14.6 | 1,058 | 11.5 | 10.8–12.2 | 23.4 | 22.2–24.7 | 558 | 6.1 | 5.6–6.6 | 35.4 | 33.0–37.8 |
| ≥ 85 | 4,811 | 109.9 | 106.8–113.1 | 19.7 | 19.2–20.2 | 966 | 22.1 | 20.7–23.5 | 29.8 | 28.3–31.5 | 521 | 11.9 | 10.9–13.0 | 41.3 | 38.5–44.0 |
| Total | 13,077 | 9.2 | 9.0–9.4 | 13.1 | 12.9–13.4 | 3.917 | 2.8 | 2.7–2.8 | 20.1 | 19.5–20.7 | 2,176 | 1.5 | 1.5–1.6 | 29.4 | 28.4–30.5 |
| Age of males (years) | |||||||||||||||
| < 1 | 147 | 9.2 | 7.8–10.9 | 8.4 | 7.2–9.8 | 46 | 2.9 | 2.1–3.9 | 9.6 | 7.1–12.7 | 49 | 3.1 | 2.3–4.1 | 38.0 | 29.6–46.9 |
| 1–14 | 16 | 0.1 | < 0.1–0.1 | 3.7 | 2.1–5.9 | 17 | 0.1 | < 0.1–0.1 | 4.5 | 2.7–7.2 | 14 | 0.1 | < 0.1–0.1 | 4.9 | 2.7–8.1 |
| 15–44 | 343 | 0.6 | 0.6–0.7 | 9.3 | 8.4–10.3 | 216 | 0.4 | 0.3–0.5 | 11.7 | 10.3–13.3 | 120 | 0.2 | 0.2–0.3 | 13.5 | 11.3–15.9 |
| 45–64 | 2,301 | 6.5 | 6.2–6.8 | 12.9 | 12.4–13.4 | 1,296 | 3.7 | 3.5–3.9 | 18.0 | 17.1–18.9 | 605 | 1.7 | 1.6–1.8 | 21.9 | 20.4–23.5 |
| 65–74 | 3,171 | 23.7 | 22.8–24.5 | 14.1 | 13.7–14.6 | 1,441 | 10.8 | 10.2–11.3 | 18.8 | 17.9–19.7 | 792 | 5.9 | 5.5–6.3 | 25.6 | 24.0–27.1 |
| 75–84 | 5,123 | 67.1 | 65.3–69.0 | 17.4 | 16.9–17.8 | 1,885 | 24.7 | 23.6–25.8 | 20.7 | 19.9–21.5 | 898 | 11.8 | 11.0–12.6 | 24.5 | 23.1–25.9 |
| ≥ 85 | 4,845 | 188.8 | 183.5–194.2 | 22.9 | 22.4–23.5 | 1,575 | 61.4 | 58.4–64.5 | 26.8 | 25.6–27.9 | 691 | 26.9 | 25.0–29.0 | 27.5 | 25.7–29.2 |
| Total | 15,946 | 11.5 | 11.3–11.6 | 16.5 | 16.2–16.7 | 6,476 | 4.7 | 4.5–4.8 | 19.9 | 19.5–20.3 | 3,169 | 2.3 | 2.2–2.4 | 23.7 | 23.0–24.5 |
| Healthcare-associationb | |||||||||||||||
| Healthcare-associated | 4,836 | 2.7 | 2.6–2.7 | 20.6 | 20.1–21.1 | 2,540 | 1.4 | 1.3–1.4 | 24.8 | 24.0–25.7 | 1,360 | 0.7 | 0.7–0.8 | 28.8 | 27.5–30.1 |
| Community-associated | 6,391 | 2.3 | 2.2–2.3 | 13.1 | 12.8–13.4 | 1,958 | 0.7 | 0.7–0.7 | 17.1 | 16.5–17.8 | 848 | 0.3 | 0.3–0.3 | 23.4 | 22.1–24.8 |
| Total | 11,227 | 4,498 | 2,208 | ||||||||||||
| Primary focus of infection | |||||||||||||||
| Urinary tract | 5,421 | 9.8 | 9.6–10.0 | 1,193 | 13.5 | 12.8–14.3 | 465 | 14.9 | 13.6–16.2 | ||||||
| Hepatobiliary | 2,739 | 14.2 | 13.7–14.7 | 916 | 18.0 | 17.0–19.1 | 159 | 30.5 | 26.5–34.6 | ||||||
| Gastrointestinal (not hepatobiliary) | 1,750 | 22.3 | 21.4–23.3 | 526 | 25.7 | 23.8–27.6 | 184 | 27.6 | 24.2–31.1 | ||||||
| Respiratory tract | 1,946 | 27.4 | 26.4–28.5 | 958 | 34.4 | 32.7–36.2 | 675 | 45.4 | 42.8–47.9 | ||||||
| Other | 789 | 14.3 | 13.4–15.2 | 362 | 13.9 | 12.6–15.3 | 397 | 19.3 | 17.6–21.1 | ||||||
| No underlying focus of infection | 536 | 19.3 | 17.8–20.8 | 184 | 24.4 | 21.4–27.7 | 130 | 34.1 | 29.4–39.1 | ||||||
| Total | 13,181 | 4,139 | 2,010 | ||||||||||||
BSI: bloodstream infection; CFR: case fatality rate; CI: confidence interval; MR: mortality rate.
a Mortality rate of healthcare-associated BSI is per 100,000 overnight bed days and day-admissions. Incidence rate of hospital-onset BSI is per overnight bed days. Every other mortality rate is per 100,000 population.
b Only data on infections between April 2020 and March 2022 were included.
Predictors of mortality
We included 65,915 (88.5%) E. coli, 19,617 (87.7%) Klebsiella spp. and 7,551 (88.0%) P. aeruginosa episodes of BSI between 1 April 2020 and 31 March 2022 in the logistic regression analysis. We found no evidence of considerable multicollinearity from the scaled generalised inflation factors for the variables included in the regression model, more details can be seen in Supplementary Table 3. After adjusting for confounders, the aOR of death in HOHA-BSI and COHA-BSI compared with CA-BSI was 2.0 (95% CI: 1.9–2.2; p < 0.01) and 1.4 (95% CI 1.4–1.5; p < 0.01) for E. coli, 2.1 (95% CI: 2.0–2.3, p < 0.01) and 1.4 (95% CI: 1.2–1.5, p < 0.01) for Klebsiella spp., and 1.7 (95% CI: 1.5–2.0, p < 0.01) and 1.3 (95% CI: 1.1–1.6, p < 0.01) for P. aeruginosa (Table 4). The odds of death with MDR-BSI compared with non-MDR-BSI were not statistically significant (E. coli: aOR = 1.1; 95% CI: 1.0–1.1; p = 0.1; Klebsiella spp.: aOR = 1.0; 95% CI: 0.9–1.1; p = 0.7 and P. aeruginosa: aOR = 1.2; 95% CI: 0.9–1.6; p = 0.2).
Table 4. Logistic regression analysis for predictors of mortality of Escherichia coli (n = 65,915), Klebsiella spp. (n = 19,617) and Pseudomonas aeruginosa (n = 7,551) bloodstream infections, England, 1 April 2020–31 March 2022.
| Characteristics | Escherichia coli | Klebsiella species | Pseudomonas aeruginosa | ||||||
|---|---|---|---|---|---|---|---|---|---|
| aOR | 95% CI | p value | aOR | 95% CI | p value | aOR | 95% CI | p value | |
| Age (years) | |||||||||
| 1–14 | 1.0 | Reference | |||||||
| < 1 | 2.8 | 1.4–6.4 | < 0.01 | 3.6 | 1.6–9.6 | < 0.01 | 9.6 | 3.9–27.4 | < 0.01 |
| 15–44 | 2.7 | 1.4–6.0 | < 0.01 | 4.5 | 2.1–11.5 | < 0.01 | 4.5 | 2.1–11.8 | < 0.01 |
| 45–64 | 6.3 | 3.3–13.9 | < 0.01 | 9.3 | 4.5–23.6 | < 0.01 | 9.4 | 4.5–24.1 | < 0.01 |
| 65–74 | 7.4 | 3.9–16.2 | < 0.01 | 11.3 | 5.5–28.8 | < 0.01 | 13.9 | 6.7–35.5 | < 0.01 |
| 75–84 | 9.1 | 4.8–20.1 | < 0.01 | 12.2 | 5.9–31.1 | < 0.01 | 15.1 | 7.3–38.8 | < 0.01 |
| ≥ 85 | 12.8 | 6.8–28.2 | < 0.01 | 19.5 | 9.4–49.8 | < 0.01 | 18.7 | 8.9–47.9 | < 0.01 |
| Sex | |||||||||
| Female | 1.0 | Reference | |||||||
| Male | 1.2 | 1.1–1.3 | < 0.01 | 1.0 | 0.9–1.1 | 1.0 | 0.8 | 0.7–0.8 | < 0.01 |
| Multidrug resistance | |||||||||
| MDR | 1.1 | 1.0–1.1 | 0.1 | 1.0 | 0.9–1.1 | 0.7 | 1.2 | 0.9–1.6 | 0.2 |
| Healthcare-association | |||||||||
| CA | 1.0 | Reference | |||||||
| COHA | 1.4 | 1.4–1.5 | < 0.01 | 1.4 | 1.2–1.5 | < 0.01 | 1.2 | 1.1–1.4 | < 0.01 |
| HOHA | 2.0 | 1.9–2.2 | < 0.01 | 2.1 | 2.0–2.3 | < 0.01 | 1.7 | 1.5–2.0 | < 0.01 |
aOR: adjusted odds ratio; CA: community-associated; CI: confidence interval; COHA: community-onset healthcare-associated; HOHA: hospital-onset healthcare-associated; MDR: multidrug resistance.
Discussion
We described the epidemiology of the three most common GNBSI in England: UKHSA routinely compares GNBSI reports with data on isolates from UKHSA’s Second Generation Surveillance System – a voluntary laboratory-based surveillance scheme, to monitor and follow up on missed cases. Between April 2019 and March 2020, it was estimated that 89%, 97%, 94% of all E. coli, Klebsiella spp. and P. aeruginosa BSI cases in England were notified to the HCAI DCS [17]. Most cases were captured by the surveillance scheme and therefore included in our study. We considered the likelihood of sampling bias as low because of the mandatory requirement to notify these GNBSI episodes.
Escherichia coli was the most common cause of BSI. The annual incidence of E. coli BSI declined during the study period, mostly between April 2019 and March 2021 during the COVID-19 pandemic. Cancelled or delayed elective procedures during the pandemic [18] likely contributed to this reduction but investigations to better understand this reduction are ongoing. Between 2016 and 2018, fewer E. coli BSI cases were notified [17]. The exact cause of this is not yet known but increased awareness of GNBSIs following the introduction of the UK’s 2016 ambition to half the number of HA-GNBSI may have improved infection control practices. Conversely, incidence of all Klebsiella spp. increased in the period 2017–22. Sloot et al. (2022) showed that during the pandemic, HOHA-BSI cases of Klebsiella spp. and P. aeruginosa were also positive for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) [19].
Like in other studies, most of the infections in our study were CA [9,20], most older patients were associated with UTI [20,21], and HA-BSI cases had higher CFR [21].
The epidemiology of the three infections differs, pointing towards the need for separate approaches and focus areas to decrease the infections. A greater proportion of E. coli BSI episodes were CA compared with Klebsiella spp. and Pseudomonas aeruginosa. The proportion of CA Klebsiella spp. and P. aeruginosa BSI has increased over time in the UK [21,22]. In the studies of Livermore et al. 2008a and Livermore et al. 2008b, 39.6% Klebsiella spp. and 33% of P. aeruginosa BSI were community-onset BSI (combination of CA-BSI and COHA-BSI) compared with 52.6% and 43.2% being CA-BSI in our study. Although there were differences between these studies, this change could demonstrate the importance of non-hospital settings, such as care homes or homes for these infections, as some invasive procedures are increasingly managed in these settings [23]. In a study on the use of catheters in patients receiving NHS-funded care, about a quarter of all patients with urinary catheters were from non-hospital settings [24]. In our study, about half of all Klebsiella spp. and P. aeruginosa BSI with a UTI as a primary focus of infection were in patients with a urinary catheter inserted or manipulated within a month before. Reporting information on the insertion and manipulation of a urinary catheter (as well as other predisposing risk factors) to the HCAI DCS is voluntary [11] and therefore limits the generalisability of our finding.
The CFR in P. aeruginosa BSI (25.8%; 95% CI: 25.2–26.4%) was higher than in E. coli (14.8%; 95% CI: 14.6–14.9) and Klebsiella spp. (20.0%; 95% CI: 19.6–20.3). However, the aOR in HA-BSI relative to CA-BSI was lowest in P. aeruginosa BSI cases.
Except for carbapenem resistance, resistance to the tested antimicrobial agents was greater in E. coli and Klebsiella spp. isolates compared with P. aeruginosa. Overall, resistance was higher in HA-BSI cases. This is not surprising given the higher selection pressure for resistant strains in healthcare settings due to greater antimicrobial use [25]. Interpreting the burden of AMR infections can be challenging due to factors such as local antimicrobial resistance and prescription rates, the burden of resistant strains, availability of alternative treatment options and MDR [26]. In our study, to mitigate this, MDR was used to assess burden, where it was found not to be a predictor of mortality. However, caution should be taken in interpreting this as evidence of no impact of AMR on mortality as we did not have information on appropriate antimicrobial therapy. In other studies, inappropriate antimicrobial therapy has had varying impact on the risk of mortality [27,28]. Future work will investigate the impact of AMR and appropriate prescribing on mortality and other patient outcomes. Exclusion of cases with unknown MDR result might have affected the results but we found no statistically significant difference in odds of death between MDR and non-MDR-GNBSI compared with those with unknown MDR data.
Surveillance of GNBSI is essential to developing and implementing prevention and control strategies in the UK and supporting the UK’s national action plan (NAP) for antimicrobial resistance [29]. Reducing AMR and incidence of GNBSI is one of the outcomes of the UK NAP. It is important that continued GNBSI surveillance is robust enough to provide insights on groups with the greatest burden of infection and help target interventions.
Conclusion
Bloodstream infections caused by Gram-negative bacteria are severe infections which can be fatal. Most BSIs in our study were community-associated and urinary tract infection was the most common primary focus of infection. To reduce GNBSI, appropriate management of UTIs and stringent hygienic routines in catheterisation are necessary.
Ethical statement
No ethical approval was sought for this study as UKHSA have approval under Regulation 3 of the Health Service (Control of Patient Information) Regulations 2002 to process national surveillance data, containing patient-identifiable information, without consent.
Funding statement
No specific funding was received for the study.
Use of artificial intelligence tools
None declared.
Data availability
The study team will consider granting access to anonymised data upon reasonable request.
Acknowledgements
We are grateful to colleagues in UKHSA, laboratories and infection prevention and control teams in NHS acute trusts across England. Their routine contribution is a vital part of the national surveillance schemes.
Supplementary Data
Conflict of interest: None declared.
Authors’ contributions: ON led the data collection, analysis, writing and corrections of the manuscript. ST and RH contributed to the study conception. ON, ST, RH, SG, DC and RG contributed to the interpretation of results, reviewed the manuscript, and approved the final manuscript that was submitted.
References
- 1.United Kingdom Health Security Agency (UKHSA). English surveillance programme for antimicrobial utilisation and resistance (ESPAUR) report 2022 to 2023. London: UKHSA; Nov 2023. Available from: https://www.gov.uk/government/publications/english-surveillance-programme-antimicrobial-utilisation-and-resistance-espaur-report
- 2. Perween N, Prakash SK, Siddiqui O. Multi drug resistant Klebsiella isolates in burn patients: a comparative study. J Clin Diagn Res. 2015;9(9):DC14-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Decraene V, Ghebrehewet S, Dardamissis E, Huyton R, Mortimer K, Wilkinson D, et al. An outbreak of multidrug-resistant Pseudomonas aeruginosa in a burns service in the North of England: challenges of infection prevention and control in a complex setting. J Hosp Infect. 2018;100(4):e239-45. 10.1016/j.jhin.2018.07.012 [DOI] [PubMed] [Google Scholar]
- 4. Tian L, Tan R, Chen Y, Sun J, Liu J, Qu H, et al. Epidemiology of Klebsiella pneumoniae bloodstream infections in a teaching hospital: factors related to the carbapenem resistance and patient mortality. Antimicrob Resist Infect Control. 2016;5(1):48. 10.1186/s13756-016-0145-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Naylor NR, Pouwels KB, Hope R, Green N, Henderson KL, Knight GM, et al. The health and cost burden of antibiotic resistant and susceptible Escherichia coli bacteraemia in the English hospital setting: A national retrospective cohort study. PLoS One. 2019;14(9):e0221944. 10.1371/journal.pone.0221944 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Shrestha P, Cooper BS, Coast J, Oppong R, Do Thi Thuy N, Phodha T, et al. Enumerating the economic cost of antimicrobial resistance per antibiotic consumed to inform the evaluation of interventions affecting their use. Antimicrob Resist Infect Control. 2018;7(1):98. 10.1186/s13756-018-0384-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Cassini A, Högberg LD, Plachouras D, Quattrocchi A, Hoxha A, Simonsen GS, et al. Attributable deaths and disability-adjusted life-years caused by infections with antibiotic-resistant bacteria in the EU and the European Economic Area in 2015: a population-level modelling analysis. Lancet Infect Dis. 2019;19(1):56-66. 10.1016/S1473-3099(18)30605-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Her Majesty’s Government (HM Government). Tackling antimicrobial resistance 2019 to 2024: the UK’s 5-year national action plan. London: HM Government; 24 Jan 2019. Available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/784894/UK_AMR_5_year_national_action_plan.pdf
- 9. Bou-Antoun S, Davies J, Guy R, Johnson AP, Sheridan EA, Hope RJ. Descriptive epidemiology of Escherichia coli bacteraemia in England, April 2012 to March 2014. Euro Surveill. 2016;21(35):30329. 10.2807/1560-7917.ES.2016.21.35.30329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Abernethy J, Guy R, Sheridan EA, Hopkins S, Kiernan M, Wilcox MH, et al. Epidemiology of Escherichia coli bacteraemia in England: results of an enhanced sentinel surveillance programme. J Hosp Infect. 2017;95(4):365-75. 10.1016/j.jhin.2016.12.008 [DOI] [PubMed] [Google Scholar]
- 11.United Kingdom Health Security Agency (UKHSA). Mandatory enhanced MRSA, MSSA and Gram-negative bacteraemia, and Clostridioides difficile infection surveillance Protocol version 4.4. London: UKHSA; May 2020. Available from: https://hcaidcs.phe.org.uk/ContentManagement/LinksAndAnnouncements/HCAIDCS_Mandatory_Surveillance_Protocol_v4.4.pdf
- 12.Office for National Statistics (ONS). Estimates of the population for the UK, England and Wales, Scotland and Northern Ireland. Newport: ONS; 8 Oct 2024. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland
- 13.National Health Service England (NHS England). Bed availability and occupancy data –day only. London: NHS England; 2022. Available from: https://www.england.nhs.uk/statistics/statistical-work-areas/bed-availability-and-occupancy/bed-data-day-only/
- 14.The European Committee on Antimicrobial Susceptibility Testing (EUCAST). Definitions of S, I and R. [Accessed: 24 Feb 2024]. Available from: https://www.eucast.org/newsiandr
- 15. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the Tidyverse. J Open Source Softw. 2019;4(43):1686. 10.21105/joss.01686 [DOI] [Google Scholar]
- 16.Fox J, Weisberg S. An R Companion to Applied Regression, Third edition. Thousand Oaks: Sage Publications, 2019. [Google Scholar]
- 17.United Kingdom Health Security Agency (UKHSA). Mandatory healthcare associated infection surveillance: data quality statement for April 2019 to March 2020. London; Jun 2022. Available from: https://www.gov.uk/government/publications/mandatory-healthcare-associated-infection-hcai-surveillance-data-quality-statement
- 18.The National Health Service England (NHS England). Hospital Admitted Patient Care Activity, 2022-23. London: NHS England; 21 Sep 2023. Available from: https://digital.nhs.uk/data-and-information/publications/statistical/hospital-admitted-patient-care-activity/2022-23
- 19. Sloot R, Nsonwu O, Chudasama D, Rooney G, Pearson C, Choi H, et al. Rising rates of hospital-onset Klebsiella spp. and Pseudomonas aeruginosa bacteraemia in NHS acute trusts in England: a review of national surveillance data, August 2020-February 2021. J Hosp Infect. 2022;119:175-81. 10.1016/j.jhin.2021.08.027 [DOI] [PubMed] [Google Scholar]
- 20.United Kingdom Health Security Agency (UKHSA). Annual epidemiological commentary: Gram-negative, MRSA, MSSA bacteraemia and C. difficile infections, up to and including financial year 2022 to 2023. London; 26 Sep 2024. Available from: https://www.gov.uk/government/statistics/mrsa-mssa-and-e-coli-bacteraemia-and-c-difficile-infection-annual-epidemiological-commentary
- 21. Livermore DM, Hope R, Brick G, Lillie M, Reynolds R, BSAC Working Parties on Resistance Surveillance . Non-susceptibility trends among Pseudomonas aeruginosa and other non-fermentative Gram-negative bacteria from bacteraemias in the UK and Ireland, 2001-06. J Antimicrob Chemother. 2008;62(Suppl 2):ii55-63. 10.1093/jac/dkn352 [DOI] [PubMed] [Google Scholar]
- 22. Livermore DM, Hope R, Brick G, Lillie M, Reynolds R, BSAC Working Parties on Resistance Surveillance . Non-susceptibility trends among Enterobacteriaceae from bacteraemias in the UK and Ireland, 2001-06. J Antimicrob Chemother. 2008;62(Suppl 2):ii41-54. 10.1093/jac/dkn351 [DOI] [PubMed] [Google Scholar]
- 23. Gómez-Zorrilla S, Becerra-Aparicio F, López Montesinos I, Ruiz de Gopegui E, Grau I, Pintado V, et al. A large multicenter prospective study of community-onset healthcare associated bacteremic urinary tract infections in the era of multidrug resistance: even worse than hospital acquired infections? Infect Dis Ther. 2021;10(4):2677-99. 10.1007/s40121-021-00537-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Shackley DC, Whytock C, Parry G, Clarke L, Vincent C, Harrison A, et al. Variation in the prevalence of urinary catheters: a profile of National Health Service patients in England. BMJ Open. 2017;7(6):e013842. 10.1136/bmjopen-2016-013842 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Neidhöfer C, Sib E, Neuenhoff M, Schwengers O, Dummin T, Buechler C, et al. Hospital sanitary facilities on wards with high antibiotic exposure play an important role in maintaining a reservoir of resistant pathogens, even over many years. Antimicrob Resist Infect Control. 2023;12(1):33. 10.1186/s13756-023-01236-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Laxminarayan R, Klugman KP. Communicating trends in resistance using a drug resistance index. BMJ Open. 2011;1(2):e000135-000135. 10.1136/bmjopen-2011-000135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ohnuma T, Chihara S, Costin B, Treggiari MM, Bartz RR, Raghunathan K, et al. Association of appropriate empirical antimicrobial therapy with in-hospital mortality in patients with bloodstream infections in the US. JAMA Netw Open. 2023;6(1):e2249353. 10.1001/jamanetworkopen.2022.49353 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Kadri SS, Lai YL, Warner S, Strich JR, Babiker A, Ricotta EE, et al. Inappropriate empirical antibiotic therapy for bloodstream infections based on discordant in-vitro susceptibilities: a retrospective cohort analysis of prevalence, predictors, and mortality risk in US hospitals. Lancet Infect Dis. 2021;21(2):241-51. 10.1016/S1473-3099(20)30477-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.His Majesty’s Government (HM Government). Confronting antimicrobial resistance 2024 to 2029. London: HM Government; 2024. Available from: https://assets.publishing.service.gov.uk/media/664394d9993111924d9d3465/confronting-antimicrobial-resistance-2024-to-2029.pdf
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.